T-REG: Preference Optimization with Token-Level Reward Regularization.

Wenxuan Zhou, Shujian Zhang, Lingxiao Zhao, Tao Meng

Browse the full ACL paper archive.